Legal claims defining the scope of protection, as filed with the USPTO.
1. A device for information centric network (ICN) approximate computation caching, the device comprising: processing circuitry; and a memory including instructions that, when the device is in operation, configure the processing circuitry to: receive an interest packet that includes: a feature set of first input data to a computation; a similarity function; and a similarity threshold; perform a search of a local data store using the feature set to determine an approximate computation result cached in the local data store, the approximate computation result based on second input data, to the computation, that differs from the first input data named in the interest packet, the search using the similarity function to compare the first input data in the feature set to the second input data for the approximate computation result and select the approximate computation result based on being within the similarity threshold; and return the approximate computation result to an author of the interest packet in response to the search.
2. The device of claim 1 , wherein the interest packet includes a tag separate from the feature set, the tag indicating that approximate computation results will satisfy the interest packet.
3. The device of claim 1 , wherein, to include the similarity function, the interest packet includes identification of the similarity function that is used to select the similarity function from a library of similarity functions hosted by a node.
4. The device of claim 1 , wherein the instructions further configure the processing circuitry to: receive a second interest packet prior to receiving the interest packet; create an entry in a PIT that uses a second feature set contained in the second interest packet after failing to find a response to the second interest packet in the approximate computation content store; and forward the second interest packet on the network.
5. The device of claim 4 , wherein the instructions further configure the processing circuitry to: receive a data packet that matches the entry in the PIT; and create an entry in the approximate computation content store that includes the second feature set from the PIT and the data in the data packet.
6. The device of claim 5 , wherein the approximate computation result is the entry in the approximate computation content store, the feature set and the second feature set being within a similarity threshold.
7. The device of claim 1 , wherein the similarity function is hierarchical.
8. The device of claim 7 , wherein the hierarchy is an ontology.
9. A method for information centric network (ICN) approximate computation caching, the method comprising: receiving an interest packet that includes: a feature set of first input data to a computation; a similarity function; and a similarity threshold; performing a search of a local data store using the feature set to determine an approximate computation result cached in the local data store, the approximate computation result based on second input data, to the computation, that differs from the first input data named in the interest packet, the search using the similarity function to compare the first input data in the feature set to the second input data for the approximate computation result and select the approximate computation result based on being within the similarity threshold; and returning the approximate computation result to an author of the interest packet in response to the search.
10. The method of claim 9 , wherein the interest packet includes a tag separate from the feature set, the tag indicating that approximate computation results will satisfy the interest packet.
11. The method of claim 9 , wherein, to include the similarity function, the interest packet includes identification of the similarity function that is used to select the similarity function from a library of similarity functions hosted by a node.
12. The method of claim 9 , comprising: receiving a second interest packet prior to receiving the interest packet; creating an entry in a PIT that uses a second feature set contained in the second interest packet after failing to find a response to the second interest packet in the approximate computation content store; and forwarding the second interest packet on the network.
13. The method of claim 12 , comprising: receiving a data packet that matches the entry in the PIT; and creating an entry in the approximate computation content store that includes the second feature set from the PIT and the data in the data packet.
14. The method of claim 13 , wherein the approximate computation result is the entry in the approximate computation content store, the feature set and the second feature set being within a similarity threshold.
15. The method of claim 9 , wherein the similarity function is hierarchical.
16. The method of claim 15 , wherein the hierarchy is an ontology.
17. At least one non-transitory machine-readable medium including instructions for information centric network (ICN) approximate computation caching, the instructions, when executed by processing circuitry, cause the processing circuitry to perform operations comprising: receiving an interest packet that includes: a feature set of first input data to a computation; a similarity function; and a similarity threshold; performing a search of a local data store using the feature set to determine an approximate computation result cached in the local data store, the approximate computation result based on second input data, to the computation, that differs from the first input data named in the interest packet, the search using the similarity function to compare the first input data in the feature set to the second input data for the approximate computation result and select the approximate computation result based on being within the similarity threshold; and returning the approximate computation result to an author of the interest packet in response to the search.
18. The at least one non-transitory machine-readable medium of claim 17 , wherein the interest packet includes a tag separate from the feature set, the tag indicating that approximate computation results will satisfy the interest packet.
19. The at least one non-transitory machine-readable medium of claim 17 , wherein, to include the similarity function, the interest packet includes identification of the similarity function that is used to select the similarity function from a library of similarity functions hosted by a node.
20. The at least one non-transitory machine-readable medium of claim 17 , wherein the operations comprise: receiving a second interest packet prior to receiving the interest packet; creating an entry in a PIT that uses a second feature set contained in the second interest packet after failing to find a response to the second interest packet in the approximate computation content store; and forwarding the second interest packet on the network.
21. The at least one non-transitory machine-readable medium of claim 20 , wherein the operations comprise: receiving a data packet that matches the entry in the PIT; and creating an entry in the approximate computation content store that includes the second feature set from the PIT and the data in the data packet.
22. The at least one non-transitory machine-readable medium of claim 21 , wherein the approximate computation result is the entry in the approximate computation content store, the feature set and the second feature set being within a similarity threshold.
23. The at least one non-transitory machine-readable medium of claim 17 , wherein the similarity function is hierarchical.
24. The at least one non-transitory machine-readable medium of claim 23 , wherein the hierarchy is an ontology.
Unknown
April 5, 2022
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.